Author Topic: Corona 1.3 Benchmark  (Read 254440 times)

2018-04-23, 13:09:34
Reply #180

maru

  • Corona Team
  • Active Users
  • ****
  • Posts: 12711
  • Marcin
    • View Profile
@Toorop:
You can use this automatic benchmark app: https://www.dropbox.com/s/50g6f06zloptqzg/corona%20automatic%20benchmark.zip?dl=0
Usage instructions:
-in the _run.bat file you can change the number of loops (times the benchmark will be run), you can set it to 1 to have just one run
-run the _run.bat file
-wait for the benchmark to finish
-after that, the result will be saved to CoronaBenchmarkResult.txt in the following form:

Code: [Select]
Corona Benchmark Results

2018-04-23 12:59:41
MARCIN-I7-6700
Microsoft Windows 10 Pro (10.0.16299)
Intel(R) Core(TM) i7-6700 CPU @ 3.40GHz
32704MB of physical memory

Iteration 1
Render Time [ms]: 0:05:27
Rays/Second: 1,483,750

Average
Render Time [ms]: 0:05:27
Rays/Second: 1,483,750

Median
Render Time [ms]: 0:05:27
Rays/Second: 1,483,750


Marcin Miodek | chaos-corona.com
3D Support Team Lead - Corona | contact us

2018-05-01, 23:06:00
Reply #181

Toorop

  • Users
  • *
  • Posts: 3
    • View Profile
@maru:
Thanks a bunch, perfect!

2018-05-12, 11:27:38
Reply #182

ultraex

  • Users
  • *
  • Posts: 1
    • View Profile
there is no x5470 xeon with 8 cores/8 thread....... wrong cpu and of course  the skor!!
please corrected or removed this post
thanks alot


2018-06-08, 08:29:46
Reply #183

gpz

  • Active Users
  • **
  • Posts: 79
    • View Profile
Hi
i get this score

2m 30s on a amd ryzen 7 2700 + 32 gb ram

But there is something i don't understand.

Why there are pc with same processor but less ram (like 16) than can get faster result?

2018-06-08, 08:55:45
Reply #184

diffuus

  • Active Users
  • **
  • Posts: 19
    • View Profile
If the benchmark render fits into your memory, there is no difference in 16 or 32 or more RAM. More RAM doesn't result in faster render times.
The clock speed of your RAM however will have an impact on the result. (Would be nice to see the clock speed in the table as well)
Another thing is the speed of the CPU itself. Some CPU's run at stock speeds, some are overclocked.

2018-08-06, 15:40:48
Reply #185

3DsChobo

  • Active Users
  • **
  • Posts: 29
    • View Profile
Every now and then, when new processors get released i consult the benchmark results page to see if updating our rendernodes with these new CPUs would be a good choice.

Now with the Benchmark fixed on corona core 1.3 I just wonder if the performance diffenrence between two systems still accurately transfers to the newest release of corona. And if the answer is no: will there be a corona benchmark using a newer render core in the future?

2018-08-07, 09:31:32
Reply #186

Ryuu

  • Former Corona Team Member
  • Active Users
  • **
  • Posts: 654
  • Michal
    • View Profile
If we released a new version of benchmark, there would certainly be some performance difference (I hope for the better), but the relative performance of 2 difrerent CPUs should be roughly still the same. We didn't do any specific CPU optimization for a long time, so anything which speeds up Corona on CPU A should (almost) equally speed it up on CPU B.

For this reason we don't have any immediate plans for releasing a new version of benchmark, unless we do some optimization targeting specific CPUs (or specific instruction sets) or we port Corona to a new platform.

2018-09-05, 11:20:38
Reply #187

boston.george

  • Active Users
  • **
  • Posts: 13
    • View Profile
Hi! I'm doing some test and I'd like to know if this render times are alligned with my specs.

I own a laptop Dell Inspiron 7559 (Intel Core i7 6700HQ, Nvidia GeForce GTX 960M, 16GB/1600Mhz, SSD 850EVO) and from benchmark tool I got 00:06:48.45, which is one of the worst results (first result was 00:04:51.49)

I've also converted an old scene done with V-Ray 2.x, where I used at the end almost only materials from the Corona Library and for one image of ~2K my laptop took 4 hours to clean up to 5%(+denoiser) at ~700K rays/s without "low on ram" alert.
Maybe is a conflictive scene because of GI and DOF, but I had read that normal values will be within 1,5M and 4M.

Thanks for reading
« Last Edit: 2018-09-05, 11:32:09 by boston.george »

2018-09-05, 17:37:25
Reply #188

TomG

  • Administrator
  • Active Users
  • *****
  • Posts: 5434
    • View Profile
Laptops will be generally slow, and your result does seem to be within the range available for that processor. Mine uses the same processor (reports a clock speed of 3.13 on the Freq column of the benchmark which is not that different from yours; does report a higher rays/s at 1,532,620 though) and came in at 5m 17s. There are some at the 7 or 8 minute mark for that processor too.
Tom Grimes | chaos-corona.com
Product Manager | contact us

2018-09-05, 21:19:19
Reply #189

boston.george

  • Active Users
  • **
  • Posts: 13
    • View Profile
So the same CPU in desktop workstation can perform up to 2X better?
They are the first times with a (slightly) biased solution, so to reduce render times I've to necessarily buy a dedicated workstation.

Thanks

2018-09-05, 21:31:36
Reply #190

TomG

  • Administrator
  • Active Users
  • *****
  • Posts: 5434
    • View Profile
I'm not sure you can get the same CPU in a desktop version - laptop CPUs are generally made specifically to be laptop CPUs (lower power consumption and other things to optimize them for that particular use). This means that in general laptops are not good rendering machines, compared to what you will get for the same price as a desktop.
Tom Grimes | chaos-corona.com
Product Manager | contact us

2018-12-04, 01:44:13
Reply #191

Mattschia

  • Active Users
  • **
  • Posts: 9
    • View Profile
    • Render Atelier Ltd
I've used corona benchmark a lot recently to test some workstations with OC that would be used as rendernode, I've noticed a big difference in watt consumption and temperature of cores between the benchmark and the actual render with v2/3, to the point that I dropped the benchmark completely. I assume it's because the latter uses the new AVX instrunction and a more recent version of embree, among other changes. Are there any plans to release an updated version of the benchmark? I believe it would be helpful also to show the difference between the newer CPUs.

2018-12-22, 08:09:32
Reply #192

Dimer

  • Active Users
  • **
  • Posts: 52
    • View Profile
Corona 1.3 Benchmark Finished
BTR Scene 16 passes
Intel(R) Xeon(R) Platinum 8172M CPU @ 2.60GHz (x2)
 Real CPU Frequency [GHz]: 3
Render Time: 0:00:22, Rays/sec: 21 272 300

2018-12-26, 21:33:43
Reply #193

GrizzlyAdams

  • Users
  • *
  • Posts: 1
    • View Profile
Rendering took 67.064 seconds
CoronaCore::exiting renderFrame
Unique Primitives: 3253360
Primitives with instancing: 471064508
Avg samples per pixel: 16
Avg rays per sample: 51.4926
Rays/s: 7.24491e+06
Samples/s: 140719

xeon 2696 v4
64gbram

2019-01-10, 19:18:58
Reply #194

gaganailawadi

  • Users
  • *
  • Posts: 4
    • View Profile
    • Facebook Page
Corona 1.3 Benchmark Finished
BTR Scene 16 passes
AMD Ryzen Threadripper 2950X 16-Core Processor  (x2)
 Real CPU Frequency [GHz]: 3.8
Render Time: 0:01:09, Rays/sec: 7,020,770